Center for Demography and EcologyUniversity of Wisconsin-Madison

What About the Children? The Psychological and Social Well-Being

of Multiracial Adolescents

Mary E. Campbell

Jennifer Eggerling-Boeck

CDE Working Paper No. 2003-10

“What about the Children?” The Psychological and Social Well-Being of Multiracial Adolescents1

Mary E. Campbell Department of Sociology

University of Wisconsin-Madison

Jennifer Eggerling-Boeck Department of Sociology

University of Wisconsin-Madison

June 2003

WORD COUNT (including footnotes and references): 9,871

1 Please direct all correspondence to Mary E. Campbell or Jennifer Eggerling-Boeck, Sociology Department, 1180 Observatory Drive, University of Wisconsin-Madison, Madison, WI 53706 or by email at campbell@ssc.wisc.edu or jeggerli@ssc.wisc.edu. Earlier versions of this paper were presented at the 2002 meetings of the Midwest Sociological Society in Milwaukee and the American Sociological Association in Chicago. We wish to thank Lincoln Quillian, Kerry Ann Rockquemore, Gary Sandefur, and the members of the Social Psychology Brownbag at the University of Wisconsin-Madison for their helpful comments. This research uses data from the Add Health project, a program project designed by J. Richard Udry (PI) and Peter Bearman, and funded by grant P01-HD31921 from the National Institute of Child Health and Human Development to the Carolina Population Center, University of North Carolina at Chapel Hill, with cooperative funding from 17 other agencies. Persons interested in obtaining data files from The National Longitudinal Study of Adolescent Health should contact Add Health, Carolina Population Center, 123 West Franklin Street, Chapel Hill, NC 27516-2524 (http://www.cpc.unc.edu/addhealth).

‘What about the Children?’ The Psychological and Social Well-Being of Multiracial Adolescents

Abstract

This paper uses the National Longitudinal Study of Adolescent Health (Add Health) to

examine the social and psychological well-being of multiracial adolescents. Using three

different measures of multiracial identity, we investigate the ways in which these adolescents

compare to their monoracial counterparts on the outcomes of depression, seriously considering

suicide, feeling socially accepted, feeling close to others at school, and participating in

extracurricular activities. We find that multiracial adolescents do experience some negative

outcomes, with considerable variation across different multiracial groups. We find no consistent

evidence, however, that multiracial adolescents face more difficulty in adolescence than

members of other racial and ethnic minority groups.

1

In the second half of the twentieth century the dramatic rise in the rates of interracial

marriages gave rise to a biracial baby boom (Root 1996).2 From 1970 to 1990 the number of

children living with one white parent and one black, Asian, or American Indian parent more than

tripled, rising from 400,000 to 1.5 million (Tatum 1997). In fact, more than one million first

generation biracial individuals have been born since 1989 (Root 1996).3 Root asserts, “the

contemporary presence of racially mixed people is unmatched in our country’s previous history”

(1996: xiv). Further, she notes that the number of biracial births is increasing at a faster rate than

the number of monoracial births and that this trend is similar across racial groups.

The growth of the multiracial population presents a number of intriguing questions for

sociologists studying race and ethnicity. Traditionally, both social scientists and the general

public have viewed biracial identity as marginal and inherently problematic (Thornton 1996).

Among sociologists, Robert Park was one of the first to discuss the situation of biracial

individuals. Park (1928) believed biracial individuals fit his concept of the “marginal man,”

“one whom fate has condemned to live in two societies and in two, not merely different, but

antagonistic, cultures” (in Stonequist 1937: xviii). He believed this marginal status led to social

and psychological maladjustment.

Folk wisdom has also held that multiracial status leads to psychological and social

difficulties. Although opposition to racial intermarriage has decreased dramatically in the past

2 This is not to suggest that interracial unions or biracial children are new phenomena. What is new is the movement to identify these individuals as multiracial rather than grouping them into monoracial groups. For example, in 1945, an article in the American Journal of Sociology discussed the plight of the “Mestizos of South Carolina.” According to Berry, these mixed race individuals were struggling to be classified as white, while the whites were trying to relegate them to Negro status. Some individuals chose the intermediate status of Indian, but he notes no cases where the individuals tried to assert their multiracial heritage. 3 First generation biracial individuals have parents of two different racial groups. Multigenerational multiracial individuals have a parent who is multiracial.

2

fifty years,4 it still exists and is often discussed with reference to worries about the children of

such unions (Bonilla-Silva and Hovespian 2000; Frankenberg 1993). Field (1996) believes that

the popular media has pathologized the experience of multiracial children. Many scholars have

noted a common reaction to biracial children—interview and survey respondents often assume

that biracial children won’t be accepted by any of the groups in their racial heritage and that this

will prove problematic for their self-concept and racial identity development (Field 1996;

Korgen 1998; Nakashima 1992; Tatum 1997; Thornton 1996; Tizard and Phoenix 1993).

However, the social movements of the 1960s and 1970s and the increasing rate of racial

intermarriage have changed the racial terrain of the United States.5 The assumption that

multiracial identity results in problematic mental health outcomes and poor social adjustment

must be questioned. Do multiracial individuals face a difficult situation as a result of their

multifaceted heritage? How do these individuals cope with their racial/ethnic identities and do

these identities lead to poorer mental health and social adjustment outcomes?

This paper examines the social and psychological well-being of multiracial students in

order to find out whether these youth are indeed having serious difficulties in their adolescent

years. Using the National Longitudinal Study of Adolescent Health (Add Health), we investigate

the ways in which these adolescents compare to their monoracial counterparts on the outcomes

of depression, serious consideration of suicide, feelings of social acceptance, connection to

school, and extracurricular participation.

Theoretical Mechanisms

4 The proportion of the population approving of racial intermarriage increased consistently from 1958 to 1995. By the end of this period, over 60% of Americans approved of interracial marriage, compared to less than 10% in 1958 (Schuman et al. 1997). 5 For a summary of trends in rates of interracial marriage over the past fifty years see Sandefur et al (2001).

3

A review of the literature reveals several possible mechanisms to explain why multiracial

heritage might lead to poorer mental health and social adjustment. First, Park’s “marginal man”

theory asserts that observers recognize that multiracial individuals do not fit clearly into any

monoracial category and therefore reject them. Park theorizes that this rejection leads to poor

mental health and social adjustment, since biracial individuals are denied the support of peer and

extended family networks.

A second possible mechanism deals with a discrepancy between expressed race and

observed race.6 Rockquemore and Brunsma (2001) note that some multiracial individuals

choose an “unvalidated” racial identity. In other words, the racial labels these individuals choose

for themselves are not recognized by others (except perhaps by close family and friends).

Williams (1996) notes that observers often have difficulty racially classifying multiracial

individuals, and they therefore face intrusive questions about their racial background. Similarly,

Tatum (1997) discusses the potential distress that can result when an individual’s chosen racial

label does not match his or her appearance. This disjuncture between observed and expressed

race may be stressful for those who are regularly challenged regarding their racial identity, and is

another possible mechanism leading to poor mental health and social adjustment.

Finally, multiracial individuals may struggle over their internal race. These individuals

may feel confused about who they “really are.” This confusion over racial identity is a third

possible mechanism that might predict poor psychological and social well-being for multiracial

individuals. As an extension of this, multiracial youth may also struggle with pressure to choose

a community of peers or dating partners that reflects their racial identity. Twine (1996)

6 In the terminology of Harris and Thomas, observed race refers to the race an observer assigns to the individual and expressed race refers to “the articulations of an individual’s identity through words and actions” (2002:4). The last aspect of racial identity in Harris and Thomas’s paradigm is internal race (the race the individual considers him or herself to be).

4

interviewed 25 multiracial (one black and one non-black parent) UC Berkeley students about

their dating experiences. She notes that all respondents understood the choice of a dating partner

as an expression of racial self-identification. Korgen (1998) found a similar pattern among her

respondents. While all adolescents are faced with decisions about choosing peers and dating

partners, these choices may be more problematic and consequential for multiracial youth and

may therefore lead to poor mental health and social adjustment.

The literature suggests that these mechanisms are especially likely to come into play

during adolescence. One of the primary tasks of adolescence is making sense of the self and of

one’s identity (Erikson 1950, 1980 [1959]). Multiracial heritage may add a challenging

dimension in the struggle of adolescents to answer the question “who am I?” The social

situations of adolescence may also create discomfort regarding multiracial identity. Tatum

(1997) suggests that multiracial adolescents may feel increased pressure to choose one racial

group over another. Further, dating, which often begins during adolescence, can bring about

special difficulties for multiracial individuals as they face prejudice from potential dating

partners or their parents (Korgen 1998; Twine 1996). Though racial identity may shift and

change across the life course, adolescence is a time when issues of identity are highlighted, and

is therefore a particularly important time to examine the effects of multiracial identity on mental

health and social adjustment.

Previous Studies of Multiracial Americans

Early work. As mentioned above, Robert Park (1928) considered multiracial individuals

in the United States to be a prime example of the “marginal man” concept. He described the

mixed-race individual as “one who lives in two worlds, in both of which he is more or less of a

5

stranger” (1928:893). Park envisioned the dilemma of the multiracial person to be experienced

as a relatively permanent period of crisis, resulting in spiritual instability, intensified self-

consciousness, restlessness, and malaise. In contrast, he also predicted a second and more

positive aspect of status as a “marginal man:” a broader and more cosmopolitan view of society

than that of nonmarginal persons (see Korgen 1998).

In their landmark examination of African Americans in Chicago after the Great

Migration, Drake and Cayton (1993 [1945]) revisited the topic of mixed race persons.7 They

noted that whites often voiced opposition to intermarriage by claiming life would be difficult for

the children of mixed marriages. Drake and Cayton concluded that intermarriage did not deter

couples from having children, but that these children were always considered black rather than

biracial. Further, the authors noted that although biracial children faced some degree of stigma

from the African American community and were usually not accepted by the white community,8

these children were able to adjust socially and were not doomed to be unhappy. Two

circumstances were most likely to cause difficulty, according to Drake and Cayton (1993

[1945]). The first was that the relatives of the white parent often did not accept the children.

The second difficult situation occurred if the mixed race individual attempted to associate with

white people. They conclude that “if the child of an interracial marriage wishes to be white but

cannot ‘pass,’ he may suffer severe maladjustment” (1993 [1945]:158). However, these cases

were a minority; “the majority make a successful adjustment in the Negro community, as

Negroes—or they pass completely over into the white group” (1993 [1945]:159).

7 Their text focuses on black-white biracial persons. 8 Unless, however, they ‘passed’ for white.

6

Drake and Cayton’s view of the adjustment of mixed race children departs from the

conclusions of Robert Park (though their book is dedicated to Park).9 Park maintained that most

biracial children were in a fairly permanent period of crisis with several negative outcomes.

Drake and Cayton concluded that though these children faced difficulties, most were able to

adjust successfully by embracing a single racial identity. These two views of the psychological

and social adjustment of mixed race children went unquestioned in decades of social science

research. It is only with the surge in the multiracial population in recent years that social

scientists began to investigate these positions more closely.

Self-identification Studies. Recently, scholars have renewed their interest in the

experiences of mixed-race people. One topic that has been explored is the ways in which

multiracial persons choose a racial label and identity.10 Rockquemore and Brunsma (2001)

surveyed and interviewed multiracial college students with one black and one white parent.

They found that these individuals chose a wide range of racial labels. The authors found four

patterns of self-identification for biracial individuals in their sample.

First, some chose a border identity, which “highlights an individual’s existence between

two socially distinct races as defining one’s biracialism” (Rockquemore and Brunsma 2001:42).

These students choose the label biracial; approximately 58% of respondents in their sample were

in this group. Second, those with a singular identity have a racial identity that is exclusively

black or white; about 17% of respondents fit this description.11 Next, those with a protean

identity change their racial identity according to context. These individuals use the labels black,

9 Of course, Drake and Cayton’s research was done almost twenty years after Park’s and was restricted to Chicago. Therefore, it may be the case that the empirical reality of biracial experiences had changed. 10 In the terminology of Harris and Thomas (2002), this refers to expressed race. See footnote 6 for descriptions of Harris and Thomas’s racial identity terms. 11 Thirteen percent labeled themselves African American and 4% labeled themselves white. However, Rockquemore and Brunsma’s sampling technique would have excluded most biracial individuals who labeled themselves white from the sample.

7

white, and biracial, choosing which to use according to the context. Approximately 4% of the

sample was included in this category. Finally, Rockquemore and Brunsma found that 13% of

those in their sample had a transcendent identity. These individuals believe race is a false

category and refuse any racial label.

Similarly, Harris and Sim (2002) found that multiracial adolescents identify in various

ways. Using data from the National Longitudinal Study of Adolescent Health, they found that

some of the students who identified as multiracial on the home survey identified as monoracial

on the school survey, and vice versa. Who is mixed race depended on the measure of race used:

6.8% of students were mixed race using the race question from the school survey, 3.6% of

students were mixed race using the question from the home survey, and 4.8% of students were

mixed race according to the combined racial designations of their parents.12 Further, the authors

conclude that the specific multiracial groups (e.g. black-white, Asian-white, black-American

Indian) may be quite different from one another. They suggest that a biracial American Indian

and white identity may be largely claimed by white students who have some weak connections to

American Indian heritage. Racial identity appears to be optional for those with one Asian and

one white parent. For students with one black and one white parent, evidence of the one-drop

rule remains; however, the situation for this group does appear to be departing somewhat from

this long-held rule.

In further work, Harris outlines three components of an individual’s race: internal race,

expressed race, and observed race. Internal race is “the race that an individual considers himself

or herself to be” (Harris and Thomas 2002:4). Expressed race is the racial label an individual

asserts through words or actions (e.g. the racial label checked on a survey). Finally, observed

12 However, data for the last category was available only for students living with both biological parents. Obviously, this creates a restricted sample.

8

race is the race that an observer assigns to an individual. These three components of race

measure differing aspects of an individual and should not be conflated. Most sociological work

measures expressed race because most surveys employ racial self-identification rather than

interviewer identification. However, in an intriguing study, Harris (2002) investigates observed

race. He notes that observed race differs by observer characteristics and therefore does not

always match expressed or internal race. These three components of race demonstrate one more

axis of variation within multiracial identity.

The studies summarized above suggest that multiracial identity can vary widely in

meaning and strength for those with a multiracial heritage. Further, multiracial identity among

adolescents is clearly socially constructed, quite varied across individuals, and many times very

fluid across time and situation for a particular individual. Finally, different components of race

are captured with different measurement tools. Therefore, scholars examining the multiracial

population must take care to note that this population is inherently difficult to define and may

shift over time and context.

Identity Management Studies. In recent research on the experiences of multiracial

persons, the processes of struggling with a multicultural heritage, finding social acceptance, and

choosing a romantic partner have been explored. These studies have examined both the

difficulties and rewards that are associated with a multiracial heritage.

Because the racial appearance of many mixed-race individuals is ambiguous, they are

often asked directly what racial or ethnic group they belong to. Williams’ (1996) used in-depth

interviews to investigate the ways in which biracial individuals responded to the ‘What are you?’

question. Williams finds that respondents had complex reactions to the experience of their racial

ambiguity. Most were active participants in shaping their racial/ethnic identity. Many

9

respondents felt that these encounters gave them a “social psychological platform to articulate

and proclaim their identities during the interaction, to rethink their identities and to empower

themselves long after the encounters had taken place, regardless of the interrogators’ motives or

their uses of biracial individuals’ responses” (Williams 1996:208).

Korgen (1998) also examined the ways in which multiracial individuals managed their

racial identities. She interviewed 64 black/white biracial individuals and analyzed their

responses to see whether they manifest the characteristics of marginality outlined by Park. She

found important differences between respondents born before 1965 and those born after this

time. Respondents born before the Civil Rights era felt it was not an option to claim a biracial

identity. Most of these respondents claimed a black identity and found acceptance in the African

American community. Therefore, biracial individuals born before the Civil Rights Movement

felt a great deal of ambiguity toward their biracial heritage. However, the great majority of her

under-thirty respondents maintained and valued a biracial identity. Most “recognize both

positive and negative consequences to having both a black and white parent, but all but two

believe the positive aspects outweigh the negative” (Korgen 1998:72).

The major difficulty voiced by Korgen’s respondents was the feeling of being an outsider

or not “fitting in.” One respondent explained that, “I sort of have defined myself as not part of

things—as sort of something else—because I’ve never really been in a group or in an

environment where I’m with other people whose parents are mixed like mine. So I guess I’ve

always felt sort of like an outsider in that sense” (Korgen 1998:75). This reflects the description

of marginality offered by Park; however, these respondents did not feel a sense of permanent

crisis, but rather temporary struggles that they were able to resolve.

10

In a more positive light, many interviewees believed that they could not be racist or could

not understand racism, due to their dual identity. One young woman reflected that, “I think that

having both a black and white parent has given me the best of both worlds. I think it’s given me

the ability to accept people for who they are and not look for black or white or use race to

discriminate or make judgments of people. It’s definitely been positive. I truly feel blessed to be

mixed” (Korgen 1998:78). This mirrors Park’s second aspect of marginality, a broader and more

cosmopolitan worldview. Korgen concludes that young biracial individuals believe the positive

aspects of being mixed race outweigh the negative aspects.

These recent studies of the mixed-race population suggest that multiracial individuals still

experience some degree of marginality, at times feeling rejected by all groups and at times

feeling forced to choose between two or more groups. However, the positive aspect of

marginality, having a more objective or broader perspective, is emphasized much more than the

negative aspects for today’s multiracial youth. How do these findings affect the mental health

and social adjustment of multiracial individuals? Does reality reflect the common assumption

that the difficulties of multiracial identity will lead individuals to be socially ostracized and less

mentally healthy? Are the benefits of a multiracial heritage that multiracial youth perceive

manifested in good mental health and social adjustment?

Mental Health and Social Adjustment Studies. A few studies have directly investigated

the mental health of multiracial youth. Based on a sample of 31 biracial (black/white) youth and

comparison groups of 31 African Americans and 31 whites, Field (1996) reports that the global

self-worth and specific self-concept13 scores are no worse for biracial adolescents than for their

African American or white peers. However, biracial respondents who chose a white reference

13 Specific self-concept was measured in the domains of social acceptance, physical attractiveness, and romantic appeal.

11

group orientation had more negative self-concepts than biracial respondents who chose either a

black or bicultural reference group orientation. This suggests that biracial adolescents who try to

merge into a white social environment may still face difficulties.

Johnson (1992) analyzes multiracial youth in Hawaii and finds that they are not alienated

or marginalized and have no greater risk for problem drinking. However, he notes the limited

generalizability of his study, due to the unique racial climate in Hawaii. Similarly, Mass (1992)

finds no significant differences in psychological adjustment, self-esteem, or self-concept between

a group of biracial white-Japanese respondents and a comparison group of monoracial Japanese

respondents. Finally, Cauce et al (1992) studied biracial children age 11 to 13 (17 black/white

and 5 Asian/white). They found virtually no significant differences between the biracial

respondents and a control group of 22 students of color on measures of family functioning,

parent-child relationships, peer relationships, and life stress.14

Research on the mental health and social adjustment of multiracial individuals has

increased greatly in the past decade, but it has been limited by several factors. First, most of the

studies of outcomes for multiracial children have been based on small, local samples, making it

hard to generalize the findings to a broader group of multiracial children. Further, many of these

samples were selected on the basis of a multiracial identity. Sampling in this way, although

almost impossible to avoid in small data sets, may create a biased sample because those with a

stronger biracial identity may be more likely to respond. 15

14 The only exceptions were that mothers of biracial adolescents were less restrictive than mothers of monoracial children (by maternal report) and biracial children reported that their friends had more negative reactions to problem behaviors such as skipping school, fighting, and stealing (compared to the students of color). 15 For example, some studies have advertised to solicit interviews from biracial individuals, which might draw only respondents who feel strongly about a biracial identity. Other studies have recruited only respondents who identified as a certain race on another survey.

12

Second, few of these studies have been able to examine the effects of context on the

social adjustment of multiracial youth. There is good reason to think that the social adjustment

of multiracial adolescents will be affected by their school environment as well as their family

and peer groups. Previous studies have had few measures of important context variables such as

the diversity of the school environment and the quality of youths’ relationships with their

parents. Finally, most of the work focuses on the issues facing black-white biracial children and

adolescents, with little discussion of multiracial adolescents with other backgrounds.

The National Longitudinal Study of Adolescent Health (Add Health) allows us to fill

many of these gaps in the literature. Add Health is a nationally representative sample of all

schools with students in the 7th to 12th grades in the United States. The survey allows

adolescents to self-identify with more than one racial group, using the same question used for

racial self-identification on the 2000 Census. Because of the large size of the sample, there are a

significant number of multiracial adolescents in the survey, allowing us to compare a reasonable

sample of multiracial adolescents to monoracial adolescents of all racial and ethnic groups. Add

Health also avoids many of the common sample selection problems that can occur with small

multiracial studies, since it does not recruit sample members based on any racial characteristics,

unlike most studies of multiracial adolescents. Thus, this study avoids the possibility that the

samples are biased toward selecting adolescents who self-identify in a particular way or feel

particularly strongly about their identification. Finally, the Add Health survey includes an

extensive set of measures, allowing us to consider many outcomes for adolescents, as well as the

effects of school and family factors.

13

Data & Methods

We examine the social adjustment and mental health of monoracial and multiracial

adolescents using the school, home and parent surveys of the National Longitudinal Study of

Adolescent Health (Add Health). Add Health surveyed a nationally representative sample of 7th

to 12th graders in the United States in 1995, completing in-home interviews with 20,745

adolescents.16 In addition, 17,700 parents of responding adolescents were surveyed, and school

administrators were surveyed about the nature of the schools attended by adolescents in the

sample. Most of the adolescents interviewed in their homes were also part of a larger survey at

school.

Independent Variables. In both the school and home surveys, students were allowed to

self-identify with more than one racial group.17 On the school survey, 1,252 of the 20,745

students identified with more than one racial group. In the home survey, 1,038 respondents

identified themselves as multiracial.18 A parent living with the respondent also completed a race

question and was allowed to mark more than one racial group for each parent’s race. We present

models using three different measures of multiracial identity.

First, we speculate that respondents who identify as multiracial both on the school survey

and at the home interview likely have a stronger or less ambiguous multiracial identity than those

who only identify as multiracial in one of the two contexts. Therefore, we create a set of dummy

variables that indicate whether an adolescent identified as multiracial in both contexts, at school

only, at home only, or in neither context. We call this measure school v. home. For respondents

16 The mean age of respondents was 15.4 years; the range was 10-21 years. 17 On the home survey respondents were asked “What is your race? You may give more than one answer.” On the school survey the question was “What is your race? If you are of more than one race, you may choose more than one.” 18 For an examination of why students might be self-identifying differently in these different contexts, see Harris and Sim (2002).

14

who were monoracial in both contexts, we identify the race they chose on the home survey.19

The first set of models tests for differences in social and psychological well-being between these

groups of multiracial students and monoracial students.

Second, in order to investigate variation between groups with different ancestries, we

present models that examine the four largest biracial detailed heritage groups on the survey:

black-white, American Indian-white, Asian-white, and black-American Indian. We use the

response to the race question on the home survey to create these multiracial groups. Finally, we

run similar detailed heritage group models in which multiracial adolescents are identified based

on the racial identification of their parents. We call this measure parental racial combination.

For these last models, we consider respondents multiracial if their parents are of two different

race groups or if at least one parent marks more than one race (we exclude students who do not

live with both biological parents). For all of the models, we exclude Hispanics, since we agree

with earlier research (such as Harris and Sim 2002) that it would be unclear whether or not

individuals who chose a Hispanic identity and a single racial identity might be thinking of

themselves as multiracial, since Add Health asks for Hispanic and racial identity in separate

questions.

Two of our sets of race variables (school v. home and the detailed heritage groups)

measure the expressed race of the adolescent (see footnote 6). Therefore, in the analyses using

these variables, those who have a monoracial expressed race are located in one of the monoracial

categories, even though their parents may be of two different races or one parent may consider

him or herself to be multiracial. Further, respondents with a multiracial expressed race are

considered multiracial, despite the fact that they might not have a multiracial heritage according

to their parent’s racial responses. The third set of race variables (parental racial combination) 19 For our sample, using the racial identification from the school survey would result in more missing cases.

15

captures expressed ancestry (a combination of the expressed race of the parents). In analyses

using this set of race variables, those who do not have parents of two different races or one

parent who considers him or herself multiracial are included in one of the monoracial categories,

even though they may have a multiracial expressed race. Further, a respondent with parents of

two different races or one parent who is multiracial will be considered multiracial even though

they may have a monoracial expressed race. Therefore each measurement of multiracial identity

leaves out some respondents who in some contexts might be considered multiracial and includes

some respondents who in some contexts might be considered monoracial. See Figure 1 to see

how these definitions overlap and the sample sizes for each definition.

[FIGURE 1 ABOUT HERE]

Dependent Variables. In order to examine the mental health and social adjustment of

adolescents, we examine several dependent variables. The first two dependent variables are

intended to measure whether or not multiracial students experience negative psychological

outcomes as a result of their unique situations. The first is a modified version of the CES-D, a

20-item depression scale. Using the 19 questions available on the AddHealth survey, a logged

depression score is computed.20 Second, we consider a dummy variable that indicates whether or

not an individual reported thinking seriously about committing suicide in the last 12 months.

The depression variable allows us to measure a low level of psychological difficulty21 while the

variable on thoughts of suicide allows us to measure a higher level of psychological difficulty.

20 The Center for Epidemiologic Studies Depression Scale is a 20-item scale with an alpha>.85. The scale we have created includes 18 of those items, as well as a question not asked on the CES-D (How often “You felt life was not worth living”). The score for each question ranges from 0 to 3, with a high score indicating a high level of depression. These items are summed and logged. According to the CES-D, about one-fourth of sampled students are depressed. For more information, see Radloff (1977). 21 We do not treat the depression variable as a threshold for clinical depression levels, but rather as a continuous variable that can capture differences in lower levels of depression among respondents.

16

Next, we analyze the self-perceived social acceptance of multiracial adolescents. One of

the main tenets of Park’s “marginal man” theory and public attitudes toward multiracial children

is that they will feel rejected by both (or all) groups in their racial heritage. Therefore our next

outcome measures to what degree individuals feel that they are socially accepted. Respondents

were asked whether they agreed or disagreed (on a five-point scale) with the statement “I feel

socially accepted.” We create a dummy variable with a value of 1 if they strongly agree or agree

with this statement.

Finally, we consider whether or not students feel a strong connection to their school. For

adolescents, the school atmosphere is one of the most important spaces for social acceptance or

rejection. Therefore, these variables are designed as indicators of whether or not multiracial

children feel isolated and rejected in the school setting. The first dependent variable used to

examine this question is a measure of connection to school. This variable is the mean of the

responses on two questions: the first asks how “close” the respondent feels to people at school,

and the second asks if the respondent feels “part of” their school. Both questions are answered

on a five-point scale. The second dependent variable in this category is a dummy variable

indicating whether or not the respondent participated in any clubs, sports teams or other

organized activities at school.

The control variables in this analysis are drawn from the school, home, and parent

surveys. Variables from the home survey include gender, age, and number of siblings. Region is

based on the location of the school. The percentage of students in the school who identify as

multiracial and the diversity of the school (measured by the percentage of students in the school

who are of a different race than the modal racial group) are calculated from the race responses of

all the students in the school. Whether or not the adolescent perceives that students at school are

17

prejudiced (coded as a 1 if the student agrees or strongly agrees that students at their school are

prejudiced) comes from the school survey as well. Socioeconomic status (measured by the

mother’s level of education and the natural log of household income)22 is constructed from

responses on the parental survey (supplemented with student responses to minimize missing

data). Parents get along with student indicates if the parents say they “always” or “often” get

along with the student. Multiracial parents indicates students who have parents of different

races. We will discuss the implications of these variables separately for each dependent variable

in the results section. Ordinary least squares regression will be used for the continuous

dependent variables (the depression scale and feeling close to one’s school), while logistic

regression will be used for the dichotomous dependent variables (having considered suicide,

feeling accepted, and participating in any clubs or activities). All models presented include

controls for design effects (for more details see Chantala and Tabor 1999).

Results

Descriptive Results [TABLE 1 ABOUT HERE]

Table 1 shows descriptive statistics for all three definitions of the key race variables and

the dependent variables. Many more adolescents self-identified as multiracial at school than at

home, and only 343 identified as multiracial in both contexts. There is also considerable

variation in the makeup of the multiracial population when defined by student report and when

defined by parent report, as was clear in Figure 1. The “black-American Indian” group is, for

example, more than twice as large if we use student self-report to define groups than it is if we

22 Mother’s level of education is grouped into 6 categories: less than 8 years of education, more than 8 years but less than a high school diploma, high school diploma, some postsecondary education, college degree, more than a college degree. Household income is measured in thousands of dollars.

18

use parent self-report. When the sample is restricted to non-Hispanic respondents only, as we do

in our analysis, we find few major changes. The most notable (and unsurprising) changes are

that the very large “other multiracial” group (by parent definition of multiracial identity) and the

“other race” group virtually disappear.

Table 1 also shows how the dependent variables differ for our sample. The average

depression score is relatively low, as we would expect, and a relatively small minority of

students report considering suicide. On the other hand, most students report being involved in

some school activity, most feel socially accepted, and the average student feels fairly close to

their school.

[TABLE 2 ABOUT HERE]

Table 2 illustrates the ways in which the dependent variables differ across the racial

groups. In order to make this table consistent with our regression models, we restrict the sample

to non-Hispanic respondents only. Under the first definition of multiracial identity, comparing

responses on the home and school surveys, differences between the multiracial groups appear

small. Using the second definition, which compares multiracial groups based on the home

survey alone, there is more variation across groups. The heterogeneous group “other

multiracial,” which includes all multiracial respondents who are not in the four largest ancestral

groups, appears to have negative outcomes across all five measures. Other patterns depend on

the outcome; black-white youth are the only group that have a pattern of consistently better

outcomes than the other groups. The final definition of multiracial identity, based on parent

identification, has a different pattern. Those whose parents identify as black and American

Indian23 have a pattern of more positive outcomes, while those with one black and one white

23 The results for black-white and black-American Indian youth should be treated with caution, since these categories include only 37 and 24 respondents, respectively.

19

parent have more mixed outcomes. The single-race respondents, also measured using the student

response to the home survey, are included to provide comparison groups.

Regression Results

[TABLE 3 ABOUT HERE]

Table 3 investigates the relationships between the first measure of multiracial identity

(school v. home identification) and these dependent variables. The first model for each

dependent variable includes only the race variables, so we can consider the racial patterns in the

outcomes. The second model includes controls that might help explain these racial differences.

The first models for each outcome show us that identifying as multiracial consistently across

both contexts (which we hypothesize is a measure of “strong” multiracial identification) is

associated with negative outcomes for the measures of mental well-being, but is not significantly

associated with the measures of social integration. Identifying as multiracial only at school is

associated with negative outcomes for all variables except club participation, while identifying as

multiracial only on the home survey has a positive association with depression and a negative

association with feeling socially accepted. Compared to white students, then, it appears that

multiracial identification is associated with a range of negative psychological and social

outcomes. It is important to note, however, that with all of these results, we do not find that

multiracial adolescents appear to bear some unique burden; instead, the monoracial minority

groups also experience various negative outcomes.

The control variables introduced in the second set of models have important relationships

with the mental health and social outcomes.24 Male students generally experience less

24 We also considered models without the school context variables, the control for getting along with parents, and the interactions. In these models, the only major difference was that those who always identify as multiracial did not have a significantly higher likelihood of joining clubs. We interpret this to mean that this positive relationship only

20

depression and thoughts of suicide, and feel more socially accepted than female students. Older

students, on the other hand, appear to have more negative mental and social outcomes than

younger students. Having parents who are highly educated and having high family income is

associated with more positive outcomes, and students who get along well with their parents (by

parental report) also experience more positive outcomes. Another family factor included in these

models is living with multiracial parents. Those whose parents self-identified as different races

were more likely to consider suicide and feel that they were not socially accepted, but the other

outcomes were unrelated to this family context variable.

School factors that we investigated include an indicator of whether or not the respondent

felt that students at the school were prejudiced. Claims that the student body was prejudiced

were associated with negative mental and social outcomes. Two other school variables were

measured directly from school data: the school diversity level and the percent of the school that

identified as multiracial. These variables had little effect; schools with larger multiracial

populations tended to have students who considered suicide more and felt less close to their

school, while school diversity was only negatively associated with club participation.

Interactions were included in the model to help clarify whether those who consistently

identified as multiracial or those who identified as multiracial only at school had different

outcomes in schools with a larger multiracial population. These interactions had little effect, but

a few relationships were observed. Those who identify as multiracial only at school tend to feel

more accepted and closer to others at their school when there is a larger multiracial population at

school, as we might expect. Those who consistently identified as multiracial, however, were less

exists in schools with very small multiracial populations, since the interaction with the size of the multiracial population is significant and negative. Other differences in the main effects were minor.

21

likely to participate in clubs and activities at a school with a larger multiracial population than

they were in a school with a very small multiracial population.

The conclusions from Table 3 are limited, however, because the multiracial categories we

are considering in this table cluster so many different ancestral groups together. There is every

reason to think that those who self-identify as black-white, for example, would have different

outcomes than those who identify as Asian-white, since the historical treatment of these groups

has been so different. The next two tables, therefore, consider two different ways of measuring

the multiracial population while breaking it down into its largest multiracial groups.

[TABLE 4 ABOUT HERE]

Table 4 shows the four largest detailed heritage groups in our sample, defined by the

self-identification of adolescents on the home survey. Those who identified as black-white on

the home survey did not differ significantly from whites on any of the outcomes. Although they

appear different from blacks on several outcomes, these differences did not reach statistical

significance when blacks were used as the comparison group. Identifying as Asian-white was

not associated with any significant differences from whites either.

Identifying as American Indian-white on the home survey was associated with higher

levels of depression than whites, but lower levels than American Indians. American Indian-

whites were slightly more likely to say they had considered suicide than whites or American

Indians, and were less likely to feel socially accepted than whites or American Indians. They

were not significantly different from whites or American Indians on the school outcomes.

Finally, those who identified as black-American Indian had depression levels that fell between

those of Blacks and American Indians, while they were less likely than either group to consider

suicide. Black-American Indians also were less likely to feel socially accepted than blacks or

22

American Indians, although they did not differ significantly compared to whites. Identifying as

Black-American Indian was also associated with a more distant connection to people at school

than Blacks or American Indians, and a lower level of club participation than either group

(although not significantly different from whites’ participation).

The second models add controls to the basic model. Inclusion of these variables changes

the multiracial relationships little, although some of the variables become insignificant as their

standard errors rise. The control variables in Table 4 are the same as those in Table 3, and most

of the associations remain the same. Attending a school with a larger multiracial population is

associated with a greater chance of considering suicide and fewer feelings of closeness to the

school. Respondents who think that the student body is prejudiced tend to have more negative

outcomes, while getting along with parents tends to be associated with positive outcomes.

School diversity is negatively associated with club participation, but has no significant

relationship with the other outcomes.

Although not shown in this table, we also calculated models that included an interaction

between multiracial identification and the size of the multiracial population in the school, in

order to consider whether the detailed heritage groups might have different outcomes in schools

with a larger multiracial population.25 Most of the interactions were not significant, but Black-

American Indians, who are generally unlikely to consider suicide, are slightly more likely to do

so in school with a larger multiracial population, while Asian-white adolescents, who generally

feel as socially accepted as whites, feel slightly less so in a school with a larger multiracial

population. These minimal effects lead us to believe that there is really little interaction with

self-identification and the multiracial population of the school.

25 Additionally, we considered models without the school context variables, the control for getting along with parents, and the interactions. These models varied only slightly from the final models presented and are therefore not included. All models are available from the authors upon request.

23

[TABLE 5 ABOUT HERE]

Table 5 presents a third measure of multiracial identity. For this table, the “multiracial”

population is defined as those with parents of different races (or at least one parent who identifies

as multiracial). Since this data is only available for those children who are living with both of

their parents, the sample size is significantly smaller for this table. Still, this is an important

measure to consider, since many past studies of multiracial youth have been based not on the

self-identification of the adolescent, but instead on the racial identities of the parents. It is also

important to remember that Park’s original mechanism for the effects of multiracial status was

based on the idea that having parents of different races created obstacles, not that self-

identification as multiracial was problematic.

The four largest multiracial groups, as defined by parental race, are included in table 5.

Those who have a black parent and a white parent were significantly (at the .05 level) less likely

than blacks or whites to feel that they were socially accepted. They were also less likely to

participate in activities in school, although the difference from blacks was not statistically

significant. Identification as American Indian-white is significantly associated with higher levels

of depression than whites, although levels were similar to American Indian adolescents.

American Indian-whites were also more likely to consider suicide than whites or American

Indians, and less likely to feel socially accepted than either group. Asian-white adolescents did

not differ significantly from white adolescents on any of the outcomes (making them less likely

to have high depression scores than Asian students, and more likely to consider suicide and feel

close to their school). The last detailed group, Black-American Indians, did not differ

significantly from whites on mental health outcomes, though American Indians had higher levels

of depression than whites. Black-American Indians felt less socially accepted than Blacks or

24

American Indians, but feel closer to their school than either group after controls are included in

the model. They did not differ significantly from whites in their level of club participation.26

The control variables have similar effects to those seen in the other models.

We also calculated models (not shown) with interactions between ancestry group and the

size of the school multiracial population for the outcomes in Table 5.27 American Indian-white

adolescents have slightly higher levels of depression and a greater chance of considering suicide

when in a school with a larger multiracial population. Black-American Indians, on the other

hand, have lower rates of considering suicide in schools with a larger multiracial population. No

other interactions were significant, leading us to conclude that there is little interaction between

group identification and the size of the multiracial population.

Discussion

These results offer mixed evidence about the mental health and social adjustment

outcomes of multiracial adolescents. Comparing tables 3 through 5 shows that different choices

about how to define the multiracial population lead to distinct conclusions. Our first set of

regression results, which compares school v. home self-identification, suggests that there are

some significant negative outcomes, especially for those who identified as multiracial only on

the school survey. Table 4, which includes detailed heritage groups according to student self-

identification, suggests that American Indian-white youth have some negative outcomes, black-

American Indian youth have mixed outcomes, but black-white and Asian-white youth do not

26 Again, results for black-white and black-American Indian youth (using parental racial combination) should be interpreted with caution due to small sample sizes. 27 Additionally, we considered models without the school context variables, the control for getting along with parents, and the interactions. These models varied only slightly from the final models presented and are therefore not included. All models are available form the authors upon request.

25

differ from whites. Table 5, which compares the same groups defined by parental racial

combination rather than student self-identification, shows that black-white youth have some

negative outcomes and black-American Indian youth have inconsistent findings, while American

Indian-white and Asian-white have findings similar to table 4. We suggest, therefore, that

studies that group all multiracial adolescents together will be more likely to find large negative

effects, while those that compare specific heritage groups will find that only some of the groups

have negative outcomes (and some also have positive outcomes, as we saw for black-American

Indian youth), although the specifics will vary depending on the way in which multiracial

adolescents are defined.

Returning to our discussion of the possible mechanisms that might lead to negative

outcomes for multiracial adolescents, Park believed that those with parents of different races

would be recognized by members of both racial groups as biracial, and could face rejection by

both groups. In this study, those who live with parents of different races are somewhat more

likely to consider suicide or feel they are not socially accepted. When we consider multiracial

groups defined by parental racial combination, we find significant negative effects on one or

more outcomes for three of the largest groups. These results suggest that there is some mixed

evidence supporting Park’s notion that children with parents of different races face some

difficulties compared to white students. However, many of the negative effects are within the

range of those experienced by monoracial minority groups. There is no way to confirm whether

this pattern results from rejection by the single race groups, or by another mechanism.

The other proposed mechanisms are also difficult to test with our data. It is unclear

whether or not multiracial students have ambiguous appearances or unvalidated personal

identities. We suggest, however, that these mechanisms are reasonable, since many multiracial

26

respondents in our survey were not consistent in their answers to racial questions across surveys.

We find that those who seem least certain about their identity (those who do not identify

consistently across the two different surveys) do indeed have some negative outcomes, especially

those who only identify on the school survey, which was an anonymous survey rather than a

face-to-face interview, like the home survey. Even so, those who do consistently identify as

multiracial on both surveys also have more negative psychological outcomes than white students

(although similar to other minority groups), so we think that the mechanisms that suggest

negative outcomes due to ambivalence about racial identity are not clearly supported by our data.

The evidence presented in these tables shows that multiracial adolescents do not always

feel happy and socially integrated. It also shows, however, that this situation is not very different

from the situation of other minority groups in the sample. In none of the tables do we find a

clear pattern of what Park suggested would be a “special” disadvantage of being multiracial.

Conclusion

Our findings support recent qualitative work on multiracial adolescents that finds both

benefits and drawbacks associated with a multiracial heritage. Although we do find that

multiracial adolescents, especially American Indian-whites, have some negative outcomes

compared to white students, we do not find evidence that Park’s “marginal man” hypothesis is

true today. Our results show no pervasive “special” disadvantages of being multiracial. With

regard to social and psychological well-being, multiracial students fall well within the range of

other adolescents who identify as part of a minority group. Indeed, for some outcomes and some

ancestry groups, biracial adolescents actually have more positive outcomes than whites.

27

Along with the evidence from qualitative studies, our findings suggest that the racial

terrain of the United States may be changing for multiracial youth. No longer does being

multiracial mean being isolated and rejected on the basis of a mixed race ancestry. However, it

is also evident that multiracial youth do not escape the effects of the larger racial structure in the

United States. Like other minority groups, they experience negative outcomes as adolescents

(compared to white students); these outcomes are related to issues such as the level of racial

prejudice in the environment.

While our results provide no consistent evidence to support Park’s strongly negative

predictions for biracial individuals, they also offer little evidence that multiracial youth are

facing environments free of psychological and social barriers. Similar to other minority groups,

multiracial youth have higher rates of depression and are more likely to consider suicide.

Feelings of social acceptance are also lower for some groups. Therefore, while our study does

not find pervasive negative outcomes for multiracial youth, we are not suggesting that multiracial

Americans should receive no sociological attention. Multiracial individuals are an important part

of the population, and their unique experiences deserve serious sociological analysis.

We note that any study of the multiracial population must take into account the varying

definitions or aspects of multiracial identity. Our study illustrates that results may change

depending on how respondents are racially categorized. Each method of racial categorization

(observed race, expressed race, internal race) may be equally valid, but researchers must specify

which method they use and discuss its implications for their findings.

We conclude that to fully understand the differences we found for the detailed heritage

groups, more studies of individual groups are needed. First, although researchers have

conducted many interview studies for Black-white youth, there have been relatively few for the

28

other detailed heritage groups. In-depth interviews can obtain nuanced information that is

simply not accessible through surveys. However, future studies must deal with the problem of

biased sample selection and small sample sizes, a formidable task indeed.

Second, longitudinal studies would allow scholars to understand how the experiences of

multiracial adolescents change throughout their lives, and help us to understand whether there is

a causal effect of multiracial identification on outcomes, or whether that effect might operate in

the opposite direction. Tracking individuals can help researchers to see how racial identification

and outcomes change over time would give us better insight into the real effects over the life

course. Similarly, studies like ours but with adult populations would be a valuable addition to

the literature. Though adolescence is a particularly interesting life period to investigate the

possible effects of multiracial identity, it is important to know what effects remain (or emerge) in

the adult multiracial population.

29

Home 478 206 217 137 School Parents 98 811 772 Total = 2719 Figure 1: Number of multiracial students by each survey (by

home survey, school survey or parent survey) Note: numbers will not sum to numbers in Table 1, because regressions that compare home and school surveys only include those who answered both surveys, while this figure includes those in our sample who answered only one of the surveys.

30

WORKS CITED

Berry, Brewton. 1945. “The Mestizos of South Carolina.” American Journal of Sociology 51: 34-41.

Bonilla-Silva, Eduardo and Mary Hovespian. 2000. “If two people are in love: Deconstructing

Whites’ views on interracial marriage with Blacks.” Presented at the 2000 meetings of the Southern Sociological Society.

Chantala, Kim and Joyce Tabor. 1999. “Strategies to Perform a Design-Based Analysis using the

Add Health Data.” Carolina Population Center, University of North Carolina-Chapel Hill: Add Health Working Paper.

Cauce, Ana Mari et al. 1992. “Between a rock and a hard place: Social adjustment of biracial

youth.” Pp. 207-222 in Racially mixed people in America, Maria P.P. Root (Ed.). Newbury Park: Sage Publications.

Drake, St. Clair and Horace R. Cayton. 1993 [1945]. Black metropolis: A study of Negro life in

a northern city. Chicago: The University of Chicago Press. Erikson, Erik H. 1950. Childhood and society. New York: Norton. -----. 1980 [1959]. Identity and the life cycle. New York: W.W. Norton and Company. Field, Lynda D. 1996. “Piecing together the puzzle: Self-concept and group identity in biracial

Black/White youth. Pp. 211-26 in The multiracial experience: Racial borders as the new frontier, Maria P.P. Root (Ed.). Thousand Oaks, CA: Sage Publications.

Frankenberg, Ruth. 1993. White women, race matters: The social construction of whiteness.

Minneapolis: University of Minnesota Press. Harris, David R. 2002. “In the eye of the beholder: Observed race and observer characteristics.”

PSC Research Report. Report No. 02-522. Population Studies Center at the Institute for Social Research, University of Michigan.

Harris, David R. and Jeremiah Joseph Sim. 2002. “Who is multiracial? The fluidity of racial

identity among U.S. adolescents.” American Sociological Review 67:614-27. Harris, David R. and Justin L. Thomas. 2002. “The educational costs of being multiracial:

Evidence from a national survey of adolescents.” PSC Research Report. Report No. 02-521. Population Studies Center at the Institute for Social Research, University of Michigan.

Johnson, Ronald C. 1992. “Offspring of cross-race and cross-ethnic marriages in Hawaii.” Pp.

239-49 in Racially mixed people in America, Maria P.P. Root (Ed.). Newbury Park: Sage Publications.

31

Korgen, Kathleen Odell. 1998. From black to biracial: Transforming racial identity among Americans. Westport, CT: Praeger.

Mass, Amy Iwasaki. 1992. “Interracial Japanese Americans: The best of both worlds or the end

of the Japanese American community?” Pp. 265-79 in Racially mixed people in America, Maria P.P. Root (Ed.). Newbury Park: Sage Publications.

Nakashima, Cynthia L. 1992. “An invisible monster: The creation and denial of mixed-race

people in America.” Pp. 162-78 in Racially mixed people in America, Maria P.P. Root (Ed.). Newbury Park: Sage Publications.

Park, Robert. 1928. “Human migration and the marginal man.” American Journal of Sociology

33:881-93. Radloff, L. S. 1977. “The CES-D scale: A self-report depression scale for research in the general

population.” Applied Psychological Measurement 1: 385-401. Rockquemore, Kerry Ann and David L. Brunsma. 2001. Beyond black: Biracial identity in

America. Thousand Oaks, CA: Sage Publications Root, Maria P.P. 1996. The multiracial experience: Racial borders as the new frontier.

Thousand Oaks, CA: Sage Publications. Sandefur, Gary D., Molly Martin, Jennifer Eggerling-Boeck, Susan E. Mannon, and Ann M.

Meier. 2001. “An overview of racial and ethnic demographic trends.” Pp. 40-102 in America becoming: Racial trends and their consequences, National Research Council. Neil J. Smelser, William Julius Wilson, and Faith Mitchell (Eds.). Washington, DC, National Academy Press.

Schuman, Howard, Charlotte Steeh, Lawrence Bobo, and Maria Krysan. 1997. Racial Attitudes

in America: Trends and interpretations (2nd ed.). Cambridge, MA: Harvard University Press. Stonequist, E.V. 1937. The marginal man: A study in personality and culture conflict. New

York: Russell and Russell. Tatum, Beverly Daniel. 1997. Why are all the Black kids sitting together in the cafeteria? New

York: Basic Books. Tizard, Barbara and Ann Phoenix. 1993. Black, White, or mixed race? Race and racism in the

lives of young people of mixed parentage. New York: Routledge. Thornton, Michael C. 1996. “Hidden agendas, identity theories, and multiracial people.” Pp.

101-20 in The multiracial experience: Racial borders as the new frontier, Maria P.P. Root (Ed.). Thousand Oaks, CA: Sage Publications.

32

Twine, Frances Winddance. 1996. “Heterosexual alliances: The romantic management of racial identity.” Pp. 291-304 in The multiracial experience: Racial borders as the new frontier, Maria P.P. Root (Ed.). Thousand Oaks, CA: Sage Publications.

Williams, Teresa Kay. 1996. “Race as process: Reassessing the ‘What are you?’ encounters of

biracial individuals.” Pp. 191-210 in The multiracial experience: Racial borders as the new frontier, Maria P.P. Root (Ed.). Thousand Oaks, CA: Sage Publications.

Table 1: Descriptive Statistics

Race Variables Percent N Non-Hispanic NMultiracial groupsSchool v. Homeb

Always multiracial 2.2 343 294School multiracial 5.9 908 782Home multiracial 2.9 440 311

Detailed heritage groups by student definitiona

Black-White 0.7 150 142American Indian-White 1.4 299 262Asian-White 0.7 143 125Black-American Indian 0.5 109 106Other Multiracial 1.6 337 168

Detailed groups by parental racial combinationc

Black-White 1.1 136 37American Indian-White 2.6 332 120Asian-White 1.6 201 101Black-American Indian 0.4 47 24Other Multiracial 4.0 508 89

Monoracial groupsa

White 57.7 11,954 10,455Black 21.3 4,404 4,320Asian 6.6 1,359 1,314American Indian 1.2 240 115Other 8.3 1,709 192

Dependent Variables Mean RangeDepression 11.4 (0-56)Closeness to school 3.8 (1-5)

PercentSeriously considered suicide 13.4Belonged to any clubs 84.1Felt socially accepted 84.4

a These racial categories are composed using the respondent's answer to the race question on the home survey.

b These racial categories are composed using the respondent's answer to the race questions on both the home and school survey.

c These racial categories are composed using the parent's answer to the race questions on the parent survey.

Table 2: Descriptive Statistics for Specific Racial Group Definitions

Depression Closeness Suicide Socially Accepted Clubs & ActivitiesRace Variables

Multiracial groups (Range 0-56) (Range 1-5)School v. Homeb

Always multiracial 11.89 3.64 21.2% 80.2% 84.4%School multiracial 11.73 3.65 17.9% 81.9% 85.7%Home multiracial 11.71 3.73 15.6% 80.7% 83.9%

Detailed Groups by student definitiona

Black-White 10.04 3.71 12.0% 81.7% 82.4%American Indian-White 12.18 3.65 20.0% 77.5% 85.6%Asian-White 12.73 3.71 19.4% 82.3% 82.8%Black-American Indian 12.19 3.59 10.5% 82.1% 85.2%Other Multiracial 12.46 3.53 22.2% 79.8% 84.1%

Detailed Groups by parent definitionc

Black-White 11.13 3.75 19.4% 78.4% 73.5%American Indian-White 11.25 3.67 22.2% 75.8% 83.3%Asian-White 10.97 3.76 19.0% 84.0% 82.9%Black-American Indian 8.83 4.04 8.3% 79.2% 87.0%Other Multiracial 9.85 3.84 14.6% 91.0% 95.8%

Monoracial groupsa

White 10.47 3.77 13.9% 84.6% 85.1%Black 11.86 3.73 10.5% 87.7% 87.8%Asian 13.01 3.84 14.9% 79.6% 84.3%American Indian 13.56 3.62 17.7% 84.4% 85.1%Other 11.14 3.63 14.9% 78.0% 83.3%

a These racial categories are composed using the respondent's answer to the race question on the home survey.b These racial categories are composed using the respondent's answer to the race questions on both the home and school survey.c These racial categories are composed using the parent's answer to the race questions on the parent survey.

PercentsMeans

Table 3: The effects of multiracial identification on well-being, using school v. home

Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2Race

Always Multiracial 0.229*** 0.263 0.719*** 0.712 -0.234 0.024 -0.141 0.177 0.110 1.147**(0.081) (0.171) (0.239) (0.530) (0.282) (0.564) (0.113) (0.272) (0.286) (0.462)

School Multiracial 0.109** 0.109 0.404*** 0.630** -0.403*** -0.839*** -0.141*** -0.338*** -0.096 -0.188(0.047) (0.078) (0.126) (0.303) (0.123) (0.269) (0.046) (0.101) (0.160) (0.326)

Home Multiracial 0.114* 0.079 0.107 -0.095 -0.407* -0.301 -0.096 -0.025 -0.107 -0.088(0.058) (0.066) (0.244) (0.292) (0.235) (0.237) (0.093) (0.105) (0.269) (0.304)

Black 0.171*** 0.102*** -0.210** -0.245** 0.206** 0.164 -0.077** -0.058 0.182 0.464***(0.028) (0.031) (0.089) (0.111) (0.101) (0.124) (0.036) (0.037) (0.120) (0.137)

Asian 0.251*** 0.212*** 0.068 -0.127 -0.378*** -0.297 0.005 0.110 0.107 0.401(0.058) (0.070) (0.167) (0.258) (0.141) (0.216) (0.056) (0.079) (0.173) (0.288)

American Indian 0.308** 0.286** 0.575* 0.451 0.181 0.433 -0.208 -0.152 0.217 0.476(0.128) (0.133) (0.306) (0.343) (0.446) (0.466) (0.151) (0.151) (0.377) (0.475)

Other -0.021 0.013 -0.751** -0.743* 0.030 -0.193 -0.016 -0.018 0.024 0.112(0.096) (0.087) (0.376) (0.426) (0.395) (0.401) (0.128) (0.122) (0.342) (0.398)

ControlsMale -0.131*** -0.481*** 0.296*** -0.007 0.070

(0.019) (0.087) (0.084) (0.027) (0.084)Mother's Education -0.053*** -0.047 0.034 0.031** 0.313***

(0.009) (0.036) (0.035) (0.012) (0.040)Age 0.036*** 0.050** -0.052** -0.036*** -0.093***

(0.006) (0.024) (0.020) (0.007) (0.029)Log of Family Income -0.047*** -0.010 0.065 0.042** 0.093*

(0.016) (0.049) (0.057) (0.016) (0.054)Number of siblings 0.041*** -0.037 0.050 -0.005 -0.030

(0.012) (0.055) (0.055) (0.016) (0.048)Multiracial parents 0.001 0.353* -0.277 -0.092 0.031

(0.046) (0.188) (0.183) (0.079) (0.241)Percent of school multiracial 0.005 0.034** -0.011 -0.015*** 0.005

(0.003) (0.013) (0.010) (0.004) (0.013)Parents get along with student -0.211*** -0.461*** 0.447*** 0.162*** 0.209*

(0.031) (0.119) (0.113) (0.042) (0.118)Students at school prejudiced 0.141*** 0.322*** -0.349*** -0.240*** -0.066

(0.021) (0.086) (0.082) (0.025) (0.091)School diversity (% not in majority) -0.016 -0.389 0.343 -0.050 -0.910**

(0.071) (0.265) (0.226) (0.090) (0.360)Interactions

Always*Percent multiracial -0.008 -0.024 -0.006 -0.013 -0.069**(0.013) (0.044) (0.040) (0.021) (0.027)

Schl only*Percent multiracial -0.002 -0.039 0.046* 0.021*** 0.008(0.007) (0.025) (0.026) (0.007) (0.024)

Observations 11690 10090 11942 10316 12024 10374 11875 10261 12031 10376Standard errors in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%All models control for design effects, and model 2 includes controls for region, missing income and missing parent race data.

Clubs & ActivitiesSocially AcceptedDepression Suicide Closeness

Table 4: The effects of multiracial identification on well-being, using detailed heritage groups defined by student's race

Race Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2Black-white 0.041 0.111 -0.009 0.036 -0.050 0.446 -0.038 0.074 0.115 0.132

(0.105) (0.124) (0.374) (0.501) (0.326) (0.410) (0.124) (0.150) (0.414) (0.383)American Indian-white 0.180*** 0.189*** 0.340* 0.385 -0.691*** -0.566* -0.126 0.000 0.018 0.039

(0.065) (0.066) (0.186) (0.261) (0.233) (0.288) (0.088) (0.114) (0.314) (0.325)Asian-white 0.119 -0.073 0.166 0.147 -0.108 0.396 -0.051 0.055 -0.200 -0.337

(0.134) (0.202) (0.404) (0.466) (0.458) (0.466) (0.122) (0.175) (0.407) (0.489)Black-American Indian 0.215* 0.153 -0.667 -1.123** -0.204 -0.202 -0.336** -0.297 -0.244 -0.062

(0.110) (0.133) (0.444) (0.559) (0.417) (0.505) (0.155) (0.231) (0.476) (0.478)Other multiracial 0.167* -0.022 0.597** 0.010 -0.203 0.252 -0.163 0.059 0.059 0.220

(0.087) (0.113) (0.286) (0.427) (0.319) (0.469) (0.135) (0.202) (0.387) (0.458)Black 0.181*** 0.102*** -0.268*** -0.251** 0.215** 0.171 -0.048 -0.060 0.180 0.466***

(0.028) (0.031) (0.083) (0.111) (0.095) (0.125) (0.034) (0.037) (0.120) (0.138)Asian 0.229*** 0.209*** -0.033 -0.141 -0.340*** -0.272 0.071 0.116 0.105 0.413

(0.052) (0.071) (0.140) (0.259) (0.127) (0.221) (0.050) (0.080) (0.174) (0.285)American Indian 0.283*** 0.297** 0.279 0.522 0.195 0.336 -0.120 -0.187 0.199 0.461

(0.100) (0.129) (0.267) (0.344) (0.292) (0.486) (0.139) (0.159) (0.378) (0.485)Other race -0.036 0.025 -0.548 -0.714* -0.138 -0.231 -0.146 -0.030 0.008 0.101

(0.080) (0.086) (0.383) (0.424) (0.269) (0.405) (0.132) (0.122) (0.341) (0.397)Controls

Male -0.132*** -0.486*** 0.296*** -0.009 0.070(0.019) (0.088) (0.084) (0.027) (0.084)

Mother's Education -0.053*** -0.046 0.033 0.032*** 0.315***(0.009) (0.036) (0.036) (0.012) (0.040)

Age 0.035*** 0.051** -0.051** -0.035*** -0.093***(0.006) (0.024) (0.020) (0.007) (0.029)

Log of Family Income -0.047*** -0.012 0.066 0.041** 0.093*(0.016) (0.050) (0.057) (0.016) (0.055)

Number of siblings 0.041*** -0.032 0.046 -0.005 -0.031(0.012) (0.055) (0.054) (0.016) (0.048)

Multiracial parents 0.015 0.368** -0.321* -0.109 0.026(0.047) (0.182) (0.178) (0.079) (0.246)

Percent of school multiracial 0.006* 0.031*** -0.010 -0.014*** 0.003(0.003) (0.012) (0.009) (0.004) (0.012)

Parents get along with student -0.210*** -0.454*** 0.443*** 0.158*** 0.211*(0.031) (0.119) (0.114) (0.042) (0.118)

Students at school are prejudiced 0.142*** 0.319*** -0.342*** -0.241*** -0.065(0.021) (0.087) (0.083) (0.025) (0.091)

School diversity (% not in majority) -0.013 -0.363 0.306 -0.050 -0.906**(0.071) (0.266) (0.227) (0.090) (0.362)

Observations 15228 10090 15519 10316 15646 10374 15369 10261 12031 10376Standard errors in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%All models control for design effects, and model 2 includes controls for region, missing income and missing parent race data.

Clubs & ActivitiesDepression Suicide ClosenessSocially accepted

Table 5: The effects of multiracial identification on well-being using parental racial combination

Race Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2Black-white 0.169 0.168 0.814 0.822 -0.064 -0.068 -0.846* -0.824 -0.949** -0.993**

(0.177) (0.176) (0.573) (0.664) (0.166) (0.179) (0.445) (0.603) (0.463) (0.453)American Indian-white 0.239*** 0.226*** 0.995*** 1.020*** -0.234 -0.192 -0.368 -0.171 -0.989*** -0.869***

(0.072) (0.070) (0.328) (0.358) (0.143) (0.174) (0.400) (0.408) (0.249) (0.306)Asian-white -0.143 -0.247 0.238 -0.153 0.058 0.144 -0.526 -0.441 0.521 0.394

(0.142) (0.181) (0.364) (0.429) (0.111) (0.144) (0.652) (0.578) (0.511) (0.540)Black-American Indian 0.103 -0.125 0.092 -1.775 0.188 0.501** 0.201 0.590 -1.740** -1.244*

(0.151) (0.139) (0.837) (1.114) (0.286) (0.212) (0.980) (1.000) (0.693) (0.727)Other multiracial 0.026 0.051 0.366 0.294 0.026 -0.166 1.128 1.312 0.268 -0.106

(0.084) (0.095) (0.426) (0.497) (0.155) (0.167) (0.860) (0.874) (0.514) (0.500)Black 0.151*** 0.096* -0.357** -0.395* -0.129*** -0.137*** -0.241 0.034 0.041 0.066

(0.040) (0.051) (0.166) (0.219) (0.044) (0.045) (0.221) (0.214) (0.157) (0.190)Asian 0.263*** 0.256*** -0.306 -0.195 0.056 0.081 -0.057 0.294 -0.452*** -0.302

(0.067) (0.071) (0.200) (0.337) (0.055) (0.075) (0.282) (0.393) (0.166) (0.245)American Indian 0.182 -0.378 -0.268 -0.177 -0.378* 0.918 0.170 0.556

(0.226) (0.337) (0.487) (0.269) (0.200) (1.205) (0.848) (0.933)Other race 0.006 0.161 0.081 -0.214 -0.284 0.174 1.009 0.046

(0.271) (0.288) (0.373) (0.175) (0.397) (1.084) (1.013) (1.015)Controls

Male -0.122*** -0.454*** 0.026 0.056 0.467***(0.024) (0.140) (0.039) (0.102) (0.104)

Mother's Education -0.057*** -0.065 0.039*** 0.383*** 0.080(0.011) (0.051) (0.013) (0.060) (0.053)

Age 0.030*** 0.098*** -0.035*** -0.081** -0.043(0.008) (0.033) (0.010) (0.040) (0.028)

Log of Family Income -0.058** 0.021 0.065*** 0.250*** 0.076(0.023) (0.085) (0.021) (0.079) (0.085)

Number of siblings 0.047*** 0.002 0.008 -0.076 0.067(0.017) (0.084) (0.021) (0.080) (0.083)

Percent of school multiracial 0.002 0.042*** -0.015*** 0.017 -0.017(0.004) (0.016) (0.005) (0.016) (0.014)

Parents get along with student -0.259*** -0.454** 0.190*** 0.141 0.486***(0.040) (0.201) (0.064) (0.183) (0.170)

Students at school are prejudiced 0.131*** 0.274** -0.142*** -0.045 -0.290**(0.030) (0.135) (0.032) (0.129) (0.116)

School diversity (% not in majority) 0.082 -0.436 -0.133 -0.836** 0.016(0.094) (0.499) (0.132) (0.413) (0.363)

Observations 6245 4880 6378 4991 6357 4981 5040 5022 6436 5031Standard errors in parentheses* significant at 10%; ** significant at 5%; *** significant at 1%All models control for design effects, and model 2 includes controls for region and missing income data.

Depression Suicide Closeness Clubs & ActivitiesSocially accepted

Center for Demography and EcologyUniversity of Wisconsin

1180 Observatory Drive Rm. 4412Madison, WI 53706-1393

U.S.A.608/262-2182

FAX 608/262-8400comments to: campbell@ssc.wisc.edu

requests to: cdepubs@ssc.wisc.edu